Neighborhood deprivation, supermarket availability , and BMI in low-income women across the urban-rural continuum: A multi-level analysis

Background: Neighborhood deprivation (an area-based measure of socioeconomic status) has been associated with increased risk of obesity and reduced supermarket availability in ecological and multilevel studies. Differences in supermarket availability may be an important link in the causal pathway between neighborhood deprivation and obesity. The primary aim of this study was to determine whether the availability of supermarkets mediated the relationship between neighborhood deprivation and obesity, and whether these relationships varied across the urban-rural continuum. Methods: Home residence of Kansas women who participated in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) between 2005-2007 (n=23,351) were geocoded within a geographic information system (GIS), and combined with geocoded supermarkets, grocery stores, and convenience stores (n=2,680). Neighborhood deprivation was calculated at the tract level using factor analysis of socioeconomic, educational, occupational, and household data. Availability of supermarkets, small grocery stores, and convenience stores was determined at the tract level, and by enumerating the number and types of stores for each WIC case within a 1-mile radius of their residence. Urban influence codes (USDA-ERS) were used to categorize counties as metropolitan, micropolitan, or rural. Multilevel models were run to fit iterative regression models using Maximum Likelihood estimates within SAS Proc Mixed. Results: There were no consistent differences in supermarket availability by tract deprivation across the urban-rural continuum. The availability of supermarkets within tract or 1-mile of residence did not mediate the relationship between neighborhood deprivation and obesity. Residing in a high deprivation tract was associated with a 0.937 unit increase in BMI after controlling for individual factors within metropolitan counties. In contrast, the association between tract deprivation and BMI was not linear in micropolitan counties, with residents in intermediate deprivation tracts having significantly lower BMI as compared to residents in both low and high deprivation tracts. Neighborhood deprivation was not associated with BMI in rural counties. These results suggest that the relationship between food environments, neighborhood deprivation, and obesity is complex and varies across the urban-rural continuum. Furthermore, these results suggest that policy efforts to attract large supermarkets into deprived neighborhoods and rural areas may not directly influence obesity, although they may offer other important health benefits.